
#!pip install phik


# import libraries 
import datetime
import os
import warnings
warnings.filterwarnings('ignore')

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns

from phik.report import correlation_report, plot_correlation_matrix

from sklearn.dummy import DummyRegressor
from sklearn.ensemble import RandomForestRegressor
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_absolute_error, make_scorer
from sklearn.model_selection import RandomizedSearchCV
from sklearn.model_selection import cross_val_score


# display settings
pd.options.display.max_columns = None
pd.set_option('display.max_colwidth', 100)
pd.set_option("display.precision", 4)

%matplotlib inline


# data path
path = os.path.join(os.path.dirname(os.getcwd()), '9_Project2')


# define function to read files and display information about the data
def read_file(dataset_type):
    """ Функция принимает строку - 
    тип датасета (full, train, tast).
    Читает данные, выводит основную информацию
    по файлу, возвращает датафрейм с данными"""
    
    # reading files
    filename = 'gold_recovery_' + dataset_type + '_new.csv'
    try:
        df = pd.read_csv(os.path.join(path, filename), 
                         parse_dates=['date'])
    except:
        df = pd.read_csv('/datasets/'+filename, 
                         parse_dates=['date'])
        
    # get first and last 5 rows
    display(df)
    
    # get information about the data
    print()
    print(df.info())
    
    # check for duplicates
    print()
    if df.duplicated().sum() > 0:
        print('В данных есть явные дубликаты')
    else:
        print('В данных нет явных дубликатов')
    
    # check for NaN
    if df.isna().sum().sum() > 0:
        print('В данных есть пропущенные значения')
    else:
        print('В данных нет пропущенных значений')
    
    return df


# complete data
data_full = read_file('full')

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 19439 entries, 0 to 19438
Data columns (total 87 columns):
 #   Column                                              Non-Null Count  Dtype         
---  ------                                              --------------  -----         
 0   date                                                19439 non-null  datetime64[ns]
 1   final.output.concentrate_ag                         19438 non-null  float64       
 2   final.output.concentrate_pb                         19438 non-null  float64       
 3   final.output.concentrate_sol                        19228 non-null  float64       
 4   final.output.concentrate_au                         19439 non-null  float64       
 5   final.output.recovery                               19439 non-null  float64       
 6   final.output.tail_ag                                19438 non-null  float64       
 7   final.output.tail_pb                                19338 non-null  float64       
 8   final.output.tail_sol                               19433 non-null  float64       
 9   final.output.tail_au                                19439 non-null  float64       
 10  primary_cleaner.input.sulfate                       19415 non-null  float64       
 11  primary_cleaner.input.depressant                    19402 non-null  float64       
 12  primary_cleaner.input.feed_size                     19439 non-null  float64       
 13  primary_cleaner.input.xanthate                      19335 non-null  float64       
 14  primary_cleaner.output.concentrate_ag               19439 non-null  float64       
 15  primary_cleaner.output.concentrate_pb               19323 non-null  float64       
 16  primary_cleaner.output.concentrate_sol              19069 non-null  float64       
 17  primary_cleaner.output.concentrate_au               19439 non-null  float64       
 18  primary_cleaner.output.tail_ag                      19435 non-null  float64       
 19  primary_cleaner.output.tail_pb                      19418 non-null  float64       
 20  primary_cleaner.output.tail_sol                     19377 non-null  float64       
 21  primary_cleaner.output.tail_au                      19439 non-null  float64       
 22  primary_cleaner.state.floatbank8_a_air              19435 non-null  float64       
 23  primary_cleaner.state.floatbank8_a_level            19438 non-null  float64       
 24  primary_cleaner.state.floatbank8_b_air              19435 non-null  float64       
 25  primary_cleaner.state.floatbank8_b_level            19438 non-null  float64       
 26  primary_cleaner.state.floatbank8_c_air              19437 non-null  float64       
 27  primary_cleaner.state.floatbank8_c_level            19438 non-null  float64       
 28  primary_cleaner.state.floatbank8_d_air              19436 non-null  float64       
 29  primary_cleaner.state.floatbank8_d_level            19438 non-null  float64       
 30  rougher.calculation.sulfate_to_au_concentrate       19437 non-null  float64       
 31  rougher.calculation.floatbank10_sulfate_to_au_feed  19437 non-null  float64       
 32  rougher.calculation.floatbank11_sulfate_to_au_feed  19437 non-null  float64       
 33  rougher.calculation.au_pb_ratio                     19439 non-null  float64       
 34  rougher.input.feed_ag                               19439 non-null  float64       
 35  rougher.input.feed_pb                               19339 non-null  float64       
 36  rougher.input.feed_rate                             19428 non-null  float64       
 37  rougher.input.feed_size                             19294 non-null  float64       
 38  rougher.input.feed_sol                              19340 non-null  float64       
 39  rougher.input.feed_au                               19439 non-null  float64       
 40  rougher.input.floatbank10_sulfate                   19405 non-null  float64       
 41  rougher.input.floatbank10_xanthate                  19431 non-null  float64       
 42  rougher.input.floatbank11_sulfate                   19395 non-null  float64       
 43  rougher.input.floatbank11_xanthate                  18986 non-null  float64       
 44  rougher.output.concentrate_ag                       19439 non-null  float64       
 45  rougher.output.concentrate_pb                       19439 non-null  float64       
 46  rougher.output.concentrate_sol                      19416 non-null  float64       
 47  rougher.output.concentrate_au                       19439 non-null  float64       
 48  rougher.output.recovery                             19439 non-null  float64       
 49  rougher.output.tail_ag                              19438 non-null  float64       
 50  rougher.output.tail_pb                              19439 non-null  float64       
 51  rougher.output.tail_sol                             19439 non-null  float64       
 52  rougher.output.tail_au                              19439 non-null  float64       
 53  rougher.state.floatbank10_a_air                     19438 non-null  float64       
 54  rougher.state.floatbank10_a_level                   19438 non-null  float64       
 55  rougher.state.floatbank10_b_air                     19438 non-null  float64       
 56  rougher.state.floatbank10_b_level                   19438 non-null  float64       
 57  rougher.state.floatbank10_c_air                     19438 non-null  float64       
 58  rougher.state.floatbank10_c_level                   19438 non-null  float64       
 59  rougher.state.floatbank10_d_air                     19439 non-null  float64       
 60  rougher.state.floatbank10_d_level                   19439 non-null  float64       
 61  rougher.state.floatbank10_e_air                     19003 non-null  float64       
 62  rougher.state.floatbank10_e_level                   19439 non-null  float64       
 63  rougher.state.floatbank10_f_air                     19439 non-null  float64       
 64  rougher.state.floatbank10_f_level                   19439 non-null  float64       
 65  secondary_cleaner.output.tail_ag                    19437 non-null  float64       
 66  secondary_cleaner.output.tail_pb                    19427 non-null  float64       
 67  secondary_cleaner.output.tail_sol                   17691 non-null  float64       
 68  secondary_cleaner.output.tail_au                    19439 non-null  float64       
 69  secondary_cleaner.state.floatbank2_a_air            19219 non-null  float64       
 70  secondary_cleaner.state.floatbank2_a_level          19438 non-null  float64       
 71  secondary_cleaner.state.floatbank2_b_air            19416 non-null  float64       
 72  secondary_cleaner.state.floatbank2_b_level          19438 non-null  float64       
 73  secondary_cleaner.state.floatbank3_a_air            19426 non-null  float64       
 74  secondary_cleaner.state.floatbank3_a_level          19438 non-null  float64       
 75  secondary_cleaner.state.floatbank3_b_air            19438 non-null  float64       
 76  secondary_cleaner.state.floatbank3_b_level          19438 non-null  float64       
 77  secondary_cleaner.state.floatbank4_a_air            19433 non-null  float64       
 78  secondary_cleaner.state.floatbank4_a_level          19438 non-null  float64       
 79  secondary_cleaner.state.floatbank4_b_air            19438 non-null  float64       
 80  secondary_cleaner.state.floatbank4_b_level          19438 non-null  float64       
 81  secondary_cleaner.state.floatbank5_a_air            19438 non-null  float64       
 82  secondary_cleaner.state.floatbank5_a_level          19438 non-null  float64       
 83  secondary_cleaner.state.floatbank5_b_air            19438 non-null  float64       
 84  secondary_cleaner.state.floatbank5_b_level          19438 non-null  float64       
 85  secondary_cleaner.state.floatbank6_a_air            19437 non-null  float64       
 86  secondary_cleaner.state.floatbank6_a_level          19438 non-null  float64       
dtypes: datetime64[ns](1), float64(86)
memory usage: 12.9 MB
None

В данных нет явных дубликатов
В данных есть пропущенные значения


# train data
data_train = read_file('train')

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 14149 entries, 0 to 14148
Data columns (total 87 columns):
 #   Column                                              Non-Null Count  Dtype         
---  ------                                              --------------  -----         
 0   date                                                14149 non-null  datetime64[ns]
 1   final.output.concentrate_ag                         14148 non-null  float64       
 2   final.output.concentrate_pb                         14148 non-null  float64       
 3   final.output.concentrate_sol                        13938 non-null  float64       
 4   final.output.concentrate_au                         14149 non-null  float64       
 5   final.output.recovery                               14149 non-null  float64       
 6   final.output.tail_ag                                14149 non-null  float64       
 7   final.output.tail_pb                                14049 non-null  float64       
 8   final.output.tail_sol                               14144 non-null  float64       
 9   final.output.tail_au                                14149 non-null  float64       
 10  primary_cleaner.input.sulfate                       14129 non-null  float64       
 11  primary_cleaner.input.depressant                    14117 non-null  float64       
 12  primary_cleaner.input.feed_size                     14149 non-null  float64       
 13  primary_cleaner.input.xanthate                      14049 non-null  float64       
 14  primary_cleaner.output.concentrate_ag               14149 non-null  float64       
 15  primary_cleaner.output.concentrate_pb               14063 non-null  float64       
 16  primary_cleaner.output.concentrate_sol              13863 non-null  float64       
 17  primary_cleaner.output.concentrate_au               14149 non-null  float64       
 18  primary_cleaner.output.tail_ag                      14148 non-null  float64       
 19  primary_cleaner.output.tail_pb                      14134 non-null  float64       
 20  primary_cleaner.output.tail_sol                     14103 non-null  float64       
 21  primary_cleaner.output.tail_au                      14149 non-null  float64       
 22  primary_cleaner.state.floatbank8_a_air              14145 non-null  float64       
 23  primary_cleaner.state.floatbank8_a_level            14148 non-null  float64       
 24  primary_cleaner.state.floatbank8_b_air              14145 non-null  float64       
 25  primary_cleaner.state.floatbank8_b_level            14148 non-null  float64       
 26  primary_cleaner.state.floatbank8_c_air              14147 non-null  float64       
 27  primary_cleaner.state.floatbank8_c_level            14148 non-null  float64       
 28  primary_cleaner.state.floatbank8_d_air              14146 non-null  float64       
 29  primary_cleaner.state.floatbank8_d_level            14148 non-null  float64       
 30  rougher.calculation.sulfate_to_au_concentrate       14148 non-null  float64       
 31  rougher.calculation.floatbank10_sulfate_to_au_feed  14148 non-null  float64       
 32  rougher.calculation.floatbank11_sulfate_to_au_feed  14148 non-null  float64       
 33  rougher.calculation.au_pb_ratio                     14149 non-null  float64       
 34  rougher.input.feed_ag                               14149 non-null  float64       
 35  rougher.input.feed_pb                               14049 non-null  float64       
 36  rougher.input.feed_rate                             14141 non-null  float64       
 37  rougher.input.feed_size                             14005 non-null  float64       
 38  rougher.input.feed_sol                              14071 non-null  float64       
 39  rougher.input.feed_au                               14149 non-null  float64       
 40  rougher.input.floatbank10_sulfate                   14120 non-null  float64       
 41  rougher.input.floatbank10_xanthate                  14141 non-null  float64       
 42  rougher.input.floatbank11_sulfate                   14113 non-null  float64       
 43  rougher.input.floatbank11_xanthate                  13721 non-null  float64       
 44  rougher.output.concentrate_ag                       14149 non-null  float64       
 45  rougher.output.concentrate_pb                       14149 non-null  float64       
 46  rougher.output.concentrate_sol                      14127 non-null  float64       
 47  rougher.output.concentrate_au                       14149 non-null  float64       
 48  rougher.output.recovery                             14149 non-null  float64       
 49  rougher.output.tail_ag                              14148 non-null  float64       
 50  rougher.output.tail_pb                              14149 non-null  float64       
 51  rougher.output.tail_sol                             14149 non-null  float64       
 52  rougher.output.tail_au                              14149 non-null  float64       
 53  rougher.state.floatbank10_a_air                     14148 non-null  float64       
 54  rougher.state.floatbank10_a_level                   14148 non-null  float64       
 55  rougher.state.floatbank10_b_air                     14148 non-null  float64       
 56  rougher.state.floatbank10_b_level                   14148 non-null  float64       
 57  rougher.state.floatbank10_c_air                     14148 non-null  float64       
 58  rougher.state.floatbank10_c_level                   14148 non-null  float64       
 59  rougher.state.floatbank10_d_air                     14149 non-null  float64       
 60  rougher.state.floatbank10_d_level                   14149 non-null  float64       
 61  rougher.state.floatbank10_e_air                     13713 non-null  float64       
 62  rougher.state.floatbank10_e_level                   14149 non-null  float64       
 63  rougher.state.floatbank10_f_air                     14149 non-null  float64       
 64  rougher.state.floatbank10_f_level                   14149 non-null  float64       
 65  secondary_cleaner.output.tail_ag                    14147 non-null  float64       
 66  secondary_cleaner.output.tail_pb                    14139 non-null  float64       
 67  secondary_cleaner.output.tail_sol                   12544 non-null  float64       
 68  secondary_cleaner.output.tail_au                    14149 non-null  float64       
 69  secondary_cleaner.state.floatbank2_a_air            13932 non-null  float64       
 70  secondary_cleaner.state.floatbank2_a_level          14148 non-null  float64       
 71  secondary_cleaner.state.floatbank2_b_air            14128 non-null  float64       
 72  secondary_cleaner.state.floatbank2_b_level          14148 non-null  float64       
 73  secondary_cleaner.state.floatbank3_a_air            14145 non-null  float64       
 74  secondary_cleaner.state.floatbank3_a_level          14148 non-null  float64       
 75  secondary_cleaner.state.floatbank3_b_air            14148 non-null  float64       
 76  secondary_cleaner.state.floatbank3_b_level          14148 non-null  float64       
 77  secondary_cleaner.state.floatbank4_a_air            14143 non-null  float64       
 78  secondary_cleaner.state.floatbank4_a_level          14148 non-null  float64       
 79  secondary_cleaner.state.floatbank4_b_air            14148 non-null  float64       
 80  secondary_cleaner.state.floatbank4_b_level          14148 non-null  float64       
 81  secondary_cleaner.state.floatbank5_a_air            14148 non-null  float64       
 82  secondary_cleaner.state.floatbank5_a_level          14148 non-null  float64       
 83  secondary_cleaner.state.floatbank5_b_air            14148 non-null  float64       
 84  secondary_cleaner.state.floatbank5_b_level          14148 non-null  float64       
 85  secondary_cleaner.state.floatbank6_a_air            14147 non-null  float64       
 86  secondary_cleaner.state.floatbank6_a_level          14148 non-null  float64       
dtypes: datetime64[ns](1), float64(86)
memory usage: 9.4 MB
None

В данных нет явных дубликатов
В данных есть пропущенные значения


# test data
data_test = read_file('test')

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5290 entries, 0 to 5289
Data columns (total 53 columns):
 #   Column                                      Non-Null Count  Dtype         
---  ------                                      --------------  -----         
 0   date                                        5290 non-null   datetime64[ns]
 1   primary_cleaner.input.sulfate               5286 non-null   float64       
 2   primary_cleaner.input.depressant            5285 non-null   float64       
 3   primary_cleaner.input.feed_size             5290 non-null   float64       
 4   primary_cleaner.input.xanthate              5286 non-null   float64       
 5   primary_cleaner.state.floatbank8_a_air      5290 non-null   float64       
 6   primary_cleaner.state.floatbank8_a_level    5290 non-null   float64       
 7   primary_cleaner.state.floatbank8_b_air      5290 non-null   float64       
 8   primary_cleaner.state.floatbank8_b_level    5290 non-null   float64       
 9   primary_cleaner.state.floatbank8_c_air      5290 non-null   float64       
 10  primary_cleaner.state.floatbank8_c_level    5290 non-null   float64       
 11  primary_cleaner.state.floatbank8_d_air      5290 non-null   float64       
 12  primary_cleaner.state.floatbank8_d_level    5290 non-null   float64       
 13  rougher.input.feed_ag                       5290 non-null   float64       
 14  rougher.input.feed_pb                       5290 non-null   float64       
 15  rougher.input.feed_rate                     5287 non-null   float64       
 16  rougher.input.feed_size                     5289 non-null   float64       
 17  rougher.input.feed_sol                      5269 non-null   float64       
 18  rougher.input.feed_au                       5290 non-null   float64       
 19  rougher.input.floatbank10_sulfate           5285 non-null   float64       
 20  rougher.input.floatbank10_xanthate          5290 non-null   float64       
 21  rougher.input.floatbank11_sulfate           5282 non-null   float64       
 22  rougher.input.floatbank11_xanthate          5265 non-null   float64       
 23  rougher.state.floatbank10_a_air             5290 non-null   float64       
 24  rougher.state.floatbank10_a_level           5290 non-null   float64       
 25  rougher.state.floatbank10_b_air             5290 non-null   float64       
 26  rougher.state.floatbank10_b_level           5290 non-null   float64       
 27  rougher.state.floatbank10_c_air             5290 non-null   float64       
 28  rougher.state.floatbank10_c_level           5290 non-null   float64       
 29  rougher.state.floatbank10_d_air             5290 non-null   float64       
 30  rougher.state.floatbank10_d_level           5290 non-null   float64       
 31  rougher.state.floatbank10_e_air             5290 non-null   float64       
 32  rougher.state.floatbank10_e_level           5290 non-null   float64       
 33  rougher.state.floatbank10_f_air             5290 non-null   float64       
 34  rougher.state.floatbank10_f_level           5290 non-null   float64       
 35  secondary_cleaner.state.floatbank2_a_air    5287 non-null   float64       
 36  secondary_cleaner.state.floatbank2_a_level  5290 non-null   float64       
 37  secondary_cleaner.state.floatbank2_b_air    5288 non-null   float64       
 38  secondary_cleaner.state.floatbank2_b_level  5290 non-null   float64       
 39  secondary_cleaner.state.floatbank3_a_air    5281 non-null   float64       
 40  secondary_cleaner.state.floatbank3_a_level  5290 non-null   float64       
 41  secondary_cleaner.state.floatbank3_b_air    5290 non-null   float64       
 42  secondary_cleaner.state.floatbank3_b_level  5290 non-null   float64       
 43  secondary_cleaner.state.floatbank4_a_air    5290 non-null   float64       
 44  secondary_cleaner.state.floatbank4_a_level  5290 non-null   float64       
 45  secondary_cleaner.state.floatbank4_b_air    5290 non-null   float64       
 46  secondary_cleaner.state.floatbank4_b_level  5290 non-null   float64       
 47  secondary_cleaner.state.floatbank5_a_air    5290 non-null   float64       
 48  secondary_cleaner.state.floatbank5_a_level  5290 non-null   float64       
 49  secondary_cleaner.state.floatbank5_b_air    5290 non-null   float64       
 50  secondary_cleaner.state.floatbank5_b_level  5290 non-null   float64       
 51  secondary_cleaner.state.floatbank6_a_air    5290 non-null   float64       
 52  secondary_cleaner.state.floatbank6_a_level  5290 non-null   float64       
dtypes: datetime64[ns](1), float64(52)
memory usage: 2.1 MB
None

В данных нет явных дубликатов
В данных есть пропущенные значения


# percentage of data for test
f'Процент данных для теста: {len(data_test) / len(data_full):.0%}'

'Процент данных для теста: 27%'


# train and test data
ax=data_train.plot(x='date', 
                y='rougher.input.feed_au',
                kind='scatter', c='gray', alpha=0.3, figsize=(15,5), label='Обучение')
data_test.plot(x='date', 
               y='rougher.input.feed_au',
               kind='scatter', c='c', alpha=0.3, label='Тест', ax=ax)

plt.xlabel('Дата')
plt.ylabel('% золота в руде')
plt.legend(title='Данные для модели', loc='upper left')
plt.title('Изменение доли золота в руде', fontsize=14)
plt.show()


# distribution histograms of feed sizes and au concentrations
fig, axs = plt.subplots(1, 2, figsize=(10, 4))
fig.suptitle("Сравнение выборок", fontsize="x-large")

sns.histplot(data=data_full, x='rougher.input.feed_size', kde=True, stat= 'density', 
             color="salmon", ax=axs[0])
sns.histplot(data=data_train, x='rougher.input.feed_size', kde=True, stat= 'density',
             color="skyblue", ax=axs[0])
sns.histplot(data=data_test, x='rougher.input.feed_size', kde=True, stat= 'density',
             color="green", ax=axs[0]).set(xlabel='Размер гранул сырья', 
                                           ylabel=None, xlim=(0, 150))

sns.histplot(data=data_full, x='rougher.input.feed_au', kde=True, stat= 'density',
             color="salmon", ax=axs[1])
sns.histplot(data=data_train, x='rougher.input.feed_au', kde=True, stat= 'density',
             color="skyblue", ax=axs[1])
sns.histplot(data=data_test, x='rougher.input.feed_au', kde=True, stat= 'density',
             color="green", ax=axs[1]).set(xlabel='Доля золота в руде', 
                                           ylabel=None, xlim=(0, 20))

plt.legend(labels=["Полные данные",
                   "Обучающие данные",
                   "Тестовые данные"], 
           bbox_to_anchor=(1.01, 1), loc='upper left', borderaxespad=0)
plt.show()


# features unavalable in the test data
set(data_train) - set(data_test)

{'final.output.concentrate_ag',
 'final.output.concentrate_au',
 'final.output.concentrate_pb',
 'final.output.concentrate_sol',
 'final.output.recovery',
 'final.output.tail_ag',
 'final.output.tail_au',
 'final.output.tail_pb',
 'final.output.tail_sol',
 'primary_cleaner.output.concentrate_ag',
 'primary_cleaner.output.concentrate_au',
 'primary_cleaner.output.concentrate_pb',
 'primary_cleaner.output.concentrate_sol',
 'primary_cleaner.output.tail_ag',
 'primary_cleaner.output.tail_au',
 'primary_cleaner.output.tail_pb',
 'primary_cleaner.output.tail_sol',
 'rougher.calculation.au_pb_ratio',
 'rougher.calculation.floatbank10_sulfate_to_au_feed',
 'rougher.calculation.floatbank11_sulfate_to_au_feed',
 'rougher.calculation.sulfate_to_au_concentrate',
 'rougher.output.concentrate_ag',
 'rougher.output.concentrate_au',
 'rougher.output.concentrate_pb',
 'rougher.output.concentrate_sol',
 'rougher.output.recovery',
 'rougher.output.tail_ag',
 'rougher.output.tail_au',
 'rougher.output.tail_pb',
 'rougher.output.tail_sol',
 'secondary_cleaner.output.tail_ag',
 'secondary_cleaner.output.tail_au',
 'secondary_cleaner.output.tail_pb',
 'secondary_cleaner.output.tail_sol'}


# add targets to the test dataset
data_test = data_test.merge(
    data_full[
        ['date', 'rougher.output.recovery', 'final.output.recovery']], 
    on='date', how='left')
data_test.shape

(5290, 55)


# drop "output" features from the train dataset
data_train = data_train[list(data_test)]
data_train.shape

(14149, 55)


# function for calculation of Recovery
def recovery(df, c, f, t):
    """Функция рассчитывает эффективность обогащения.
    Принимает на вход: 
    df - датафрейм; 
    с — строку с названием столбца с долей золота в концентрате 
    ПОСЛЕ флотации/очистки;
    f — строку с названием столбца с долей золота в сырье/концентрате 
    ДО флотации/очистки;
    t — строку с названием столбца с долей золота в отвальных хвостах 
    после флотации/очистки.
    Возвращает pd.Series c коэффициентами эффективности обогащения.
    """
    rec = (df[c] * (df[f]-df[t])
           .div(df[f]*(df[c]-df[t]))
           .mul(100))
    return rec


# calculation of recovery after rougher process
rec_rougher = recovery(data_full,
                       'rougher.output.concentrate_au',
                       'rougher.input.feed_au',
                       'rougher.output.tail_au')


# calculation of final Recovery
rec_final = recovery(data_full,
                     'final.output.concentrate_au',
                     'rougher.input.feed_au',
                     'final.output.tail_au')


print('Средний коэффициент эффективности обогащения:')
print()
print('Черновой концентрат:')
print(f'Рассчеты {rec_rougher.mean():.2f}')
print(f'Данные {data_full["rougher.output.recovery"].mean():.2f}')
print(f'MAE: {mean_absolute_error(data_full["rougher.output.recovery"], rec_rougher):.4f}')
print()
print('Финальный продукт:')
print(f'Рассчеты {rec_final.mean():.2f}')
print(f'Данные {data_full["final.output.recovery"].mean():.2f}')
print(f'MAE: {mean_absolute_error(data_full["final.output.recovery"], rec_final):.4f}')

Средний коэффициент эффективности обогащения:

Черновой концентрат:
Рассчеты 83.33
Данные 83.33
MAE: 0.0000

Финальный продукт:
Рассчеты 67.05
Данные 67.05
MAE: 0.0000


# boxplots
data_full[['rougher.output.recovery', 'final.output.recovery']].plot(
    kind='box', title='Коэффициенты обогащения', figsize=(10,5))
plt.ylabel('%')
plt.xticks([1,2], ['Черновой концентрат', 'Финальный продукт'])
plt.show()

#descriptive statistics
data_full[['rougher.output.recovery', 'final.output.recovery']].describe()


# boxplots
data_full[['rougher.output.concentrate_au', 
           'rougher.output.tail_au',
           'final.output.concentrate_au',
           'final.output.tail_au']].plot(
    kind='box', title='Распределения долей золота', figsize=(10,5))
plt.ylabel('%')
plt.xticks([1,2,3,4], ['Черновой концентрат', 
                       'Черновые отвалы', 
                       'Финальный продукт', 
                       'Финальные отвалы'])
plt.show()

data_full[['rougher.output.concentrate_au', 
           'rougher.output.tail_au',
           'final.output.concentrate_au',
           'final.output.tail_au']].describe()


print('Процент данных с нулевым содержанием золота в черновом концентрате:')
print(f'{len(data_full[data_full["rougher.output.concentrate_au"] < 0.001])/len(data_full):.2%}')

Процент данных с нулевым содержанием золота в черновом концентрате:
2.03%


# [Au]=0 in rougher_output
data_full[data_full['rougher.output.concentrate_au'] < 0.001][['rougher.input.feed_au', 
                                                               'rougher.output.concentrate_au',
                                                               'rougher.output.tail_au',
                                                               'rougher.output.recovery']]


# [Au]=0 in rougher_output, time distribution
ax = data_full.plot(x='date', 
                    y='rougher.input.feed_au',
                    kind='scatter', c='gray', alpha=0.2, figsize=(15,5), label='ненулевая')
data_full[data_full['rougher.output.concentrate_au'] < 0.001].plot(
    x='date', y='rougher.input.feed_au', kind='scatter', c='r', ax=ax,
    label='нулевая')

plt.xlabel('Дата')
plt.ylabel('% золота в руде')
plt.legend(title='Доля золота в черновом концентрате')
plt.title('Распределение "нулевых" долей золота в черновом концентрате')
plt.show()


print('Процент данных с нулевой концентрацией золота на финальной стадии:')
print(f'{len(data_full[data_full["final.output.concentrate_au"] < 0.001])/len(data_full):.2%}')

Процент данных с нулевой концентрацией золота на финальной стадии:
0.50%


# [Au]=0 in final.output
data_full[data_full['final.output.concentrate_au'] < 0.001][['rougher.input.feed_au', 
                                                             'final.output.concentrate_au',
                                                             'final.output.tail_au',
                                                             'rougher.output.recovery',
                                                             'final.output.recovery']]


# [Au]=0 in final.output, time distribution
ax = data_full.plot(x='date', 
                    y='rougher.input.feed_au',
                    kind='scatter', c='gray', alpha=0.2, 
                    figsize=(15,5), label='ненулевая')
data_full[data_full['final.output.concentrate_au'] < 0.001].plot(
    x='date', y='rougher.input.feed_au', kind='scatter', c='r', 
    ax=ax, label='нулевая')

plt.xlabel('Дата')
plt.ylabel('% золота в руде')
plt.legend(title='Доля золота в финальном продукте')
plt.title('Распределение "нулевых" долей золота в финальном продукте')
plt.show()


print('Процент данных с 100% обогащением, черновой концентрат:')
print(f'{len(data_full[data_full["rougher.output.recovery"] > 99.99])/len(data_full):.2%}')

Процент данных с 100% обогащением, черновой концентрат:
0.03%


# [Au]=100 in rougher_output
data_full[data_full['rougher.output.recovery'] > 99.99][['rougher.input.feed_au', 
                                                      'rougher.output.concentrate_au',
                                                      'rougher.output.tail_au',
                                                      'rougher.output.recovery']]


print('Процент данных с 100% обогащением на финальной стадии:')
print(f'{len(data_full[data_full["final.output.recovery"] > 99.99])/len(data_full):.2%}')

Процент данных с 100% обогащением на финальной стадии:
0.47%


# [Au]=100 in final.output
data_full[data_full['final.output.recovery'] > 99.99][['rougher.output.concentrate_au',
                                                      'primary_cleaner.output.concentrate_au',
                                                      'final.output.concentrate_au',
                                                      'final.output.tail_au',
                                                      'final.output.recovery']]


# [Au]=100 in final.output, time distribution
ax = data_full.plot(x='date', 
                    y='rougher.input.feed_au',
                    kind='scatter', c='gray', alpha=0.2, 
                    figsize=(15,5), label='ненулевые')
data_full[data_full['final.output.tail_au'] < 0.001].plot(
    x='date', y='rougher.input.feed_au', kind='scatter', c='r', 
    ax=ax, label='нулевые')

plt.xlabel('Дата')
plt.ylabel('% золота в руде')
plt.legend(title='Доля золота в отвальных хвостах')
plt.title('Распределение "нулевых" долей золота в финальных отвальных хвостах')
plt.show()


# scatterplots and distribution histograms, rougher
sns.pairplot(data_full[['rougher.input.feed_size',
                        'rougher.input.feed_rate',
                        'rougher.output.recovery']])

<seaborn.axisgrid.PairGrid at 0x275b2e4abe0>


# scatterplots and distribution histograms, primary_cleaner
sns.pairplot(data_full[['primary_cleaner.input.feed_size',
                        'primary_cleaner.output.concentrate_au',
                        'primary_cleaner.output.tail_au',
                        'final.output.recovery']])

<seaborn.axisgrid.PairGrid at 0x275b15edc70>


# descriptive statistics
data_full[['rougher.input.feed_size', 
           'rougher.input.feed_rate',
           'primary_cleaner.input.feed_size',
           'rougher.output.recovery',
           'final.output.recovery']].describe()


# scatterplots and distribution histograms, rougher
sns.pairplot(data_full[['rougher.input.floatbank10_sulfate',
                        'rougher.input.floatbank10_xanthate',
                        'rougher.input.floatbank11_sulfate',
                        'rougher.input.floatbank11_xanthate',
                        'rougher.output.recovery']])

<seaborn.axisgrid.PairGrid at 0x275b40d0fd0>


# scatterplot
plt.scatter(data_full['rougher.input.feed_au'], 
            data_full['rougher.output.recovery'],
            c=data_full['rougher.input.floatbank10_sulfate'],
            alpha = .5, cmap = 'viridis', 
            vmin=5, vmax=20
           )
plt.colorbar(label='Количество сульфидов')
plt.title('Черновой концентрат')
plt.xlabel('Доля золота в руде, %')
plt.ylabel('Коэффициент обогащения, %')
plt.show()


# scatterplot
plt.scatter(data_full['rougher.input.feed_au'], 
            data_full['rougher.output.recovery'],
            c=data_full['rougher.input.floatbank10_xanthate'],
            alpha = .5, cmap = 'viridis', 
            vmin=1, vmax=9
           )
plt.colorbar(label='Количество ксантогенатов')
plt.title('Черновой концентрат')
plt.xlabel('Доля золота в руде, %')
plt.ylabel('Коэффициент обогащения, %')
plt.show()


# scatterplots and distribution histograms
sns.pairplot(data_full[['rougher.input.floatbank10_sulfate',
                        'rougher.input.floatbank10_xanthate',
                        'primary_cleaner.input.sulfate',
                        'primary_cleaner.input.depressant',
                        'primary_cleaner.input.xanthate',
                        'final.output.recovery']])

<seaborn.axisgrid.PairGrid at 0x275b8ede970>


# scatterplot
plt.scatter(data_full['rougher.output.concentrate_au'], 
            data_full['final.output.recovery'],
            c=data_full['primary_cleaner.input.sulfate'],
            alpha = .5, cmap = 'viridis',
            vmin=25, vmax=250
           )
plt.colorbar(label='Количество сульфидов')
plt.title('Финальный продукт')
plt.xlabel('Доля золота в черновом концентрате, %')
plt.ylabel('Коэффициент обогащения, %')
plt.show()


# scatterplot
plt.scatter(data_full['rougher.output.concentrate_au'], 
            data_full['final.output.recovery'],
            c=data_full['primary_cleaner.input.xanthate'],
            alpha = .5, cmap = 'viridis',
            vmin=1, vmax=2
           )
plt.colorbar(label='Количество ксантогенатов')
plt.title('Финальный продукт')
plt.xlabel('Доля золота в черновом концентрате, %')
plt.ylabel('Коэффициент обогащения, %')
plt.show()


# distribution histograms, rougher
data_full[['rougher.state.floatbank10_a_air',
           'rougher.state.floatbank10_b_air',
           'rougher.state.floatbank10_c_air',
           'rougher.state.floatbank10_d_air',
           'rougher.state.floatbank10_e_air',
           'rougher.state.floatbank10_f_air']].hist(bins=100, figsize=(10, 10))
plt.show()


# distribution histograms, rougher
data_full[['rougher.state.floatbank10_a_level',
           'rougher.state.floatbank10_b_level',
           'rougher.state.floatbank10_c_level',
           'rougher.state.floatbank10_d_level',
           'rougher.state.floatbank10_e_level',
           'rougher.state.floatbank10_f_level']].hist(bins=100, figsize=(10, 10))
plt.show()


# scatterplots and distribution histograms, rougher
sns.pairplot(data_full[['rougher.state.floatbank10_b_air',
                        'rougher.state.floatbank10_b_level',
                        'rougher.output.recovery']])

<seaborn.axisgrid.PairGrid at 0x275bd831be0>


# scatterplot
plt.scatter(data_full['rougher.input.feed_au'], 
            data_full['rougher.output.recovery'],
            c=data_full['rougher.state.floatbank10_b_air'],
            alpha = .5, cmap = 'viridis')

plt.colorbar(label='объема воздуха')
plt.title('Черновой концентрат')
plt.xlabel('Доля золота в руде, %')
plt.ylabel('Коэффициент обогащения, %')
plt.show()


# scatterplot
plt.scatter(data_full['rougher.input.feed_au'], 
            data_full['rougher.output.recovery'],
            c=data_full['rougher.state.floatbank10_b_level'],
            alpha = .5, cmap = 'viridis')

plt.colorbar(label='Уровень жидкости')
plt.title('Черновой концентрат')
plt.xlabel('Доля золота в руде, %')
plt.ylabel('Коэффициент обогащения, %')
plt.show()


# scatterplots and distribution histograms, primary_cleaner
sns.pairplot(data_full[['primary_cleaner.state.floatbank8_b_air',
                        'primary_cleaner.state.floatbank8_b_level',
                        'final.output.recovery']], diag_kind='kde')

<seaborn.axisgrid.PairGrid at 0x275bef607c0>


# scatterplots and distribution histograms, secondary_cleaner
sns.pairplot(data_full[['secondary_cleaner.state.floatbank3_b_air',
                        'secondary_cleaner.state.floatbank3_b_level',
                        'final.output.recovery']], diag_kind='kde')

<seaborn.axisgrid.PairGrid at 0x275bf514a00>


# scatterplots and distribution histograms, rougher
sns.pairplot(data_full[['rougher.input.feed_au',
                        'rougher.input.feed_ag',
                        'rougher.input.feed_pb',
                        'rougher.input.feed_sol',
                        'rougher.output.recovery',
                        'final.output.recovery']])

<seaborn.axisgrid.PairGrid at 0x275bfb94880>


# descriptive statistics
data_full[['rougher.input.feed_au',
           'rougher.input.feed_ag',
           'rougher.input.feed_pb',
           'rougher.input.feed_sol']].describe()


# scatterplot
plt.scatter(data_full['rougher.input.feed_au'], 
            data_full['rougher.output.concentrate_au'],
            c=data_full['rougher.output.tail_au'],
            alpha = .5, cmap = 'viridis',
            vmin=0, vmax=8
           )
plt.colorbar(label='Отвальные хвосты')
plt.title('Доля золота, %')
plt.xlabel('Руда')
plt.ylabel('Черновой концентрат')
plt.show()


# Phik correlation coefficients matrix
phik_overview = data_full[['rougher.input.feed_au',
                       'rougher.input.feed_ag',
                       'rougher.input.feed_pb',
                       'rougher.input.feed_sol',
                       'rougher.output.recovery',
                       'final.output.recovery']].phik_matrix()

interval columns not set, guessing: ['rougher.input.feed_au', 'rougher.input.feed_ag', 'rougher.input.feed_pb', 'rougher.input.feed_sol', 'rougher.output.recovery', 'final.output.recovery']


# Plot Phik correlation coefficients matrix
plot_correlation_matrix(phik_overview.values, 
                        x_labels=('feed_au', 'feed_ag', 'feed_pb', 'feed_sol',
                                  'rougher_recovery', 'final_recovery'),
                        y_labels=('feed_au', 'feed_ag', 'feed_pb', 'feed_sol',
                                  'rougher_recovery', 'final_recovery'), 
                        title="Phik correlation coefficients", 
                        figsize=(10, 5),
                        vmin=0, vmax=1)


# distribution histograms
fig, axs = plt.subplots(2, 2, figsize=(7, 7))
fig.suptitle("Распределение долей компонентов на каждой стадии", fontsize="x-large")

sns.histplot(data=data_full, x='rougher.input.feed_au', kde=True, color="skyblue", ax=axs[0, 0])
sns.histplot(data=data_full, x='rougher.output.concentrate_au', kde=True, color="green", ax=axs[0, 0])
sns.histplot(data=data_full, x='primary_cleaner.output.concentrate_au', kde=True, color="salmon", ax=axs[0, 0])
sns.histplot(data=data_full, x='final.output.concentrate_au', kde=True, color="gold", ax=axs[0, 0]).set(
    title='Золото', xlabel='%', ylabel='Количество партий')

sns.histplot(data=data_full, x='rougher.input.feed_ag', kde=True, color="skyblue", ax=axs[0, 1])
sns.histplot(data=data_full, x='rougher.output.concentrate_ag', kde=True, color="green", ax=axs[0, 1])
sns.histplot(data=data_full, x='primary_cleaner.output.concentrate_ag', kde=True, color="salmon", ax=axs[0, 1])
sns.histplot(data=data_full, x='final.output.concentrate_ag', kde=True, color="gold", ax=axs[0, 1]).set(
    title='Серебро', xlabel='%', ylabel='Количество партий')

sns.histplot(data=data_full, x='rougher.input.feed_pb', kde=True, color="skyblue", ax=axs[1, 0])
sns.histplot(data=data_full, x='rougher.output.concentrate_pb', kde=True, color="green", ax=axs[1, 0])
sns.histplot(data=data_full, x='primary_cleaner.output.concentrate_pb', kde=True, color="salmon", ax=axs[1, 0])
sns.histplot(data=data_full, x='final.output.concentrate_pb', kde=True, color="gold", ax=axs[1, 0]).set(
    title='Свинец', xlabel='%', ylabel='Количество партий')

sns.histplot(data=data_full, x='rougher.input.feed_sol', kde=True, color="skyblue", ax=axs[1, 1])
sns.histplot(data=data_full, x='rougher.output.concentrate_sol', kde=True, color="green", ax=axs[1, 1])
sns.histplot(data=data_full, x='primary_cleaner.output.concentrate_sol', kde=True, color="salmon", ax=axs[1, 1])
sns.histplot(data=data_full, x='final.output.concentrate_sol', kde=True, color="gold", ax=axs[1, 1]).set(
    title='Зольные компоненты', xlabel='%', ylabel='Количество партий')

plt.tight_layout()
plt.legend(labels=["Руда",
                   "Черновой концентрат",
                   "Первичная очистка",
                   "Финальный концентрат"],
          bbox_to_anchor=(1.01, 1), loc='upper left', borderaxespad=0)
plt.show()


# distribution histograms
sns.histplot(data=data_full, 
             x=data_full[['rougher.input.feed_au',
                          'rougher.input.feed_ag',
                          'rougher.input.feed_pb', 
                          'rougher.input.feed_sol']].sum(axis=1), 
             kde=True, color="skyblue")

sns.histplot(data=data_full, 
             x=data_full[['rougher.output.concentrate_au',
                          'rougher.output.concentrate_ag',
                          'rougher.output.concentrate_pb',
                          'rougher.output.concentrate_sol']].sum(axis=1), 
             kde=True, color="green")

sns.histplot(data=data_full, 
             x=data_full[['primary_cleaner.output.concentrate_au',
                          'primary_cleaner.output.concentrate_ag',
                          'primary_cleaner.output.concentrate_pb',
                          'primary_cleaner.output.concentrate_sol']].sum(axis=1),
             kde=True, color="salmon")
sns.histplot(data=data_full, 
             x=data_full[['final.output.concentrate_au',
                          'final.output.concentrate_ag',
                          'final.output.concentrate_pb',
                          'final.output.concentrate_sol']].sum(axis=1),
             kde=True, color="gold").set(title='Распределение сумм долей компонентов', 
                                         xlabel='%', 
                                         ylabel='Количество партий')

plt.legend(labels=["Руда",
                   "Черновой концентрат",
                   "Первичная очистка",
                   "Финальный концентрат"])
plt.show()


anom_train=len(data_train[(data_train["rougher.output.recovery"] < 0.001) | 
               (data_train["rougher.output.recovery"] > 99.99) |
               (data_train["final.output.recovery"] < 0.001) |
               (data_train["final.output.recovery"] > 99.99)])/len(data_train)

anom_test=len(data_test[(data_test["rougher.output.recovery"] < 0.001) | 
             (data_test["rougher.output.recovery"] > 99.99) |
             (data_test["final.output.recovery"] < 0.001) |
             (data_test["final.output.recovery"] > 99.99)])/len(data_test)

print('Процент аномальных значений коэффициентов обогащения:')
print(f'Обучающая выборка {anom_train:.2%}')
print(f'Тестовая выборка {anom_test:.2%}')

Процент аномальных значений коэффициентов обогащения:
Обучающая выборка 3.02%
Тестовая выборка 2.87%


# drop 0 and 100 from the training data
data_train = data_train[(
    (data_train["rougher.output.recovery"] > 0.001) & (data_train["rougher.output.recovery"] < 99.99)) & (
    (data_train["final.output.recovery"] > 0.001) & (data_train["final.output.recovery"] < 99.99))]

print('Нулевые или =100% коэффициенты обогащения в обучающей выборке:')
print(len(data_train[(data_train["rougher.output.recovery"] < 0.001) | 
               (data_train["rougher.output.recovery"] > 99.99) |
               (data_train["final.output.recovery"] < 0.001) |
               (data_train["final.output.recovery"] > 99.99)]))

Нулевые или =100% коэффициенты обогащения в обучающей выборке:
0


# number of NaN
print(f'Количество пропущенных значений в train (N={len(data_train)})')
print(data_train.isna().sum()[data_train.isna().sum() > 0]
      .sort_values(ascending=False), '\n')

print(f'Количество пропущенных значений в test (N={len(data_test)})')
print(data_test.isna().sum()[data_test.isna().sum() > 0]
      .sort_values(ascending=False))

Количество пропущенных значений в train (N=13721)
rougher.input.floatbank11_xanthate            374
rougher.state.floatbank10_e_air               368
secondary_cleaner.state.floatbank2_a_air      211
rougher.input.feed_size                       139
rougher.input.feed_pb                         100
primary_cleaner.input.xanthate                 87
rougher.input.feed_sol                         58
rougher.input.floatbank11_sulfate              29
rougher.input.floatbank10_sulfate              19
primary_cleaner.input.depressant               15
rougher.input.feed_rate                         8
secondary_cleaner.state.floatbank2_b_air        7
primary_cleaner.input.sulfate                   7
secondary_cleaner.state.floatbank3_a_air        4
secondary_cleaner.state.floatbank4_a_air        4
rougher.input.floatbank10_xanthate              3
primary_cleaner.state.floatbank8_d_air          3
primary_cleaner.state.floatbank8_a_air          2
primary_cleaner.state.floatbank8_c_air          2
primary_cleaner.state.floatbank8_b_air          2
secondary_cleaner.state.floatbank3_b_air        1
secondary_cleaner.state.floatbank3_b_level      1
secondary_cleaner.state.floatbank6_a_air        1
secondary_cleaner.state.floatbank5_a_level      1
secondary_cleaner.state.floatbank4_a_level      1
secondary_cleaner.state.floatbank4_b_air        1
secondary_cleaner.state.floatbank4_b_level      1
secondary_cleaner.state.floatbank5_a_air        1
secondary_cleaner.state.floatbank5_b_air        1
secondary_cleaner.state.floatbank3_a_level      1
secondary_cleaner.state.floatbank5_b_level      1
rougher.state.floatbank10_b_level               1
secondary_cleaner.state.floatbank2_b_level      1
secondary_cleaner.state.floatbank2_a_level      1
rougher.state.floatbank10_c_level               1
rougher.state.floatbank10_c_air                 1
rougher.state.floatbank10_b_air                 1
rougher.state.floatbank10_a_level               1
rougher.state.floatbank10_a_air                 1
primary_cleaner.state.floatbank8_d_level        1
primary_cleaner.state.floatbank8_c_level        1
primary_cleaner.state.floatbank8_b_level        1
primary_cleaner.state.floatbank8_a_level        1
secondary_cleaner.state.floatbank6_a_level      1
dtype: int64 

Количество пропущенных значений в test (N=5290)
rougher.input.floatbank11_xanthate          25
rougher.input.feed_sol                      21
secondary_cleaner.state.floatbank3_a_air     9
rougher.input.floatbank11_sulfate            8
primary_cleaner.input.depressant             5
rougher.input.floatbank10_sulfate            5
primary_cleaner.input.sulfate                4
primary_cleaner.input.xanthate               4
rougher.input.feed_rate                      3
secondary_cleaner.state.floatbank2_a_air     3
secondary_cleaner.state.floatbank2_b_air     2
rougher.input.feed_size                      1
dtype: int64


# distribution of NaNs in time
ax = data_full.plot(x='date', 
                    y='rougher.input.feed_au',
                    kind='scatter', c='gray', alpha=0.2, 
                    figsize=(15,5), label='not NAN')
data_full[data_full['rougher.input.feed_size'].isna()].plot(
    x='date', y='rougher.input.feed_au', kind='scatter', c='r', 
    ax=ax, label='NAN, rougher.input.feed_size')
data_full[data_full['rougher.input.feed_pb'].isna()].plot(
    x='date', y='rougher.input.feed_au', kind='scatter', c='c', 
    ax=ax, label='NAN, rougher.input.feed_pb')
data_full[data_full['rougher.input.feed_sol'].isna()].plot(
    x='date', y='rougher.input.feed_au', kind='scatter', c='b', 
    ax=ax, label='NAN, rougher.input.feed_sol')

plt.xlabel('Дата')
plt.ylabel('Доля золота в руде, %')
plt.legend(title='Пропущенные значения')
plt.title('Распределение пропущенных значений в некоторых признаках')
plt.show()


# fill NAN in all dataset
data_full = data_full.fillna(method='ffill')
print('Количество пропусков в полном датасете:',
      data_full.isna().sum()[data_full.isna().sum() > 0])

Количество пропусков в полном датасете: Series([], dtype: int64)


# fill NAN in training and test data
data_train = (data_train
              .set_index('date')
              .fillna(data_full.set_index('date')))

data_test = (data_test
             .set_index('date')
             .fillna(data_full.set_index('date')))


print('Количество пропусков в обучающих данных:', 
      data_train.isna().sum()[data_train.isna().sum() > 0])
print('Количество пропусков в тестовых данных:', 
      data_test.isna().sum()[data_test.isna().sum() > 0])

Количество пропусков в обучающих данных: Series([], dtype: int64)
Количество пропусков в тестовых данных: Series([], dtype: int64)


# define function to calculate Symmetric Mean Absolute Percentage Error (sMAPE)
def smape_score(y_true, y_pred):
    y_true = np.array(y_true)
    res = np.mean(
        np.abs(y_true - y_pred) / ((np.abs(y_true) + np.abs(y_pred))*0.5)
    )*100
    return res


# define function to calculate final sMAPE
def final_smape(smape_rougher, smape_final):
    return (smape_rougher*0.25 + smape_final*0.75)


# training data for "Rougher recovery" model
data_train_rougher = data_train[['rougher.input.feed_pb',
                                 'rougher.input.feed_rate',
                                 'rougher.input.feed_size',
                                 'rougher.input.feed_sol',
                                 'rougher.input.feed_au',
                                 'rougher.input.floatbank10_sulfate',
                                 'rougher.state.floatbank10_b_air',
                                 'rougher.state.floatbank10_b_level',
                                 'rougher.output.recovery']]
data_train_rougher.shape

(13721, 9)


# test data for "Rougher recovery" model
data_test_rougher = data_test[list(data_train_rougher)]
data_test_rougher.shape

(5290, 9)


# targets and features
x_train_rougher = data_train_rougher.drop('rougher.output.recovery', axis=1)
y_train_rougher = data_train_rougher['rougher.output.recovery']

x_test_rougher = data_test_rougher.drop('rougher.output.recovery', axis=1)
y_test_rougher = data_test_rougher['rougher.output.recovery']


# Create the model
model_lr = LinearRegression()

# Cross-validation
scores_lr = cross_val_score(model_lr, x_train_rougher, y_train_rougher, cv=5, 
                            scoring=make_scorer(smape_score))


score_lr = scores_lr.mean()
print(f'Среднee sMAPE: {score_lr:.2f} со стандартным отклонением {scores_lr.std():.2f}')

Среднee sMAPE: 6.73 со стандартным отклонением 1.73


# Create the model
model_rf = RandomForestRegressor(random_state=123)

# Hyperparameters to check
params = {'max_depth': range(1, 30), 'n_estimators': range(10, 200, 10)}


# Cross-validation
grid_model_rougher = RandomizedSearchCV(model_rf, params, 
                                        n_jobs=-1, 
                                        random_state=123, cv=5,
                                        scoring=make_scorer(smape_score, greater_is_better=False))

# Fit the model on the training data
grid_model_rougher.fit(x_train_rougher, y_train_rougher)

print(f'Гиперпараметры лучшей модели: {grid_model_rougher.best_params_}')
print(f'sMAPE: {abs(grid_model_rougher.best_score_):.2f}')

Гиперпараметры лучшей модели: {'n_estimators': 190, 'max_depth': 8}
sMAPE: 6.54


# Create the dummy model
dummy_regr = DummyRegressor(strategy="mean")

# Cross-validation
scores_dr = cross_val_score(dummy_regr, x_train_rougher, y_train_rougher, cv=5, 
                            scoring=make_scorer(smape_score, greater_is_better=False))

score_dr = scores_dr.mean()

print(f'sMAPE случайной модели: {abs(score_dr):.2f} со стандартным отклонением {scores_dr.std():.2f}')

sMAPE случайной модели: 7.80 со стандартным отклонением 1.70


# training data for "Final recovery" model
data_train_final = data_train[['primary_cleaner.input.sulfate',
                               'primary_cleaner.input.xanthate',
                               'rougher.input.feed_pb',
                               'rougher.input.feed_size',
                               'rougher.input.feed_sol',
                               'rougher.input.feed_au',
                               'rougher.input.floatbank10_sulfate',
                               'secondary_cleaner.state.floatbank3_b_air',
                               'secondary_cleaner.state.floatbank3_b_level',
                               'final.output.recovery']]
data_train_final.shape

(13721, 10)


# test data for "Final recovery" model
data_test_final = data_test[list(data_train_final)]
data_test_final.shape

(5290, 10)


# targets and features
x_train_final = data_train_final.drop('final.output.recovery', axis=1)
y_train_final = data_train_final['final.output.recovery']

x_test_final = data_test_final.drop('final.output.recovery', axis=1)
y_test_final = data_test_final['final.output.recovery']


# Create the model
model_lr = LinearRegression()

# Cross-validation
scores_lr = cross_val_score(model_lr, x_train_final, y_train_final, cv=5,
                            scoring=make_scorer(smape_score, greater_is_better=False))


score_lr = scores_lr.mean()
print(f'Средняя оценка качества модели: {abs(score_lr):.2f} со стандартным отклонением {scores_lr.std():.2f}')

Средняя оценка качества модели: 9.46 со стандартным отклонением 1.51


# Create the model
model_rf = RandomForestRegressor(random_state=123)

# Hyperparameters to check
params = {'max_depth': range(1, 30), 'n_estimators': range(10, 200, 10)}


# Cross-validation
grid_model = RandomizedSearchCV(model_rf, params, n_jobs=-1, random_state=123, cv=5,
                                scoring=make_scorer(smape_score, greater_is_better=False))

# Fit the model on the training data
grid_model.fit(x_train_final, y_train_final)

print(f'Гиперпараметры лучшей модели: {grid_model.best_params_}')
print(f'sMAPE: {abs(grid_model.best_score_):.2f}')

Гиперпараметры лучшей модели: {'n_estimators': 70, 'max_depth': 5}
sMAPE: 9.95


# Create the dummy model
dummy_regr = DummyRegressor(strategy="mean")

# Cross-validation
scores_dr = cross_val_score(dummy_regr, x_train_final, y_train_final, cv=5, 
                            scoring=make_scorer(smape_score, greater_is_better=False))

score_dr = scores_dr.mean()

print(f'sMAPE случайной модели: {abs(score_dr):.2f} со стандартным отклонением {scores_dr.std():.2f}')

sMAPE случайной модели: 10.30 со стандартным отклонением 0.94


# Create the model
model_rougher = RandomForestRegressor(max_depth=8, n_estimators=190, random_state=12345)

# Fit the model on the training data
model_rougher.fit(x_train_rougher, y_train_rougher)

RandomForestRegressor(max_depth=8, n_estimators=190, random_state=12345)

RandomForestRegressor(max_depth=8, n_estimators=190, random_state=12345)


# Make predictions on the test data
pred_test_rougher = model_rougher.predict(x_test_rougher)


# Create the dummy model
dummy_regr = DummyRegressor(strategy="mean")
# Fit the model on the training data
dummy_regr.fit(x_train_rougher, y_train_rougher)
# Make predictions on the test data
pred_dummy_rougher = dummy_regr.predict(x_test_rougher)


# Evaluate the model
smape_model_rougher = smape_score(y_test_rougher, pred_test_rougher)
smape_dummy_rougher = smape_score(y_test_rougher, pred_dummy_rougher)

print('sMAPE:', round(smape_model_rougher, 4))
print('sMAPE случайной модели:', round(smape_dummy_rougher, 4))

sMAPE: 8.1025
sMAPE случайной модели: 9.0755


# Importance of the features for RandomForest
feature_importances = pd.DataFrame(
    {'features': list(x_train_rougher.columns),
     'feature_importance': model_rougher.feature_importances_})

# plot feature importances
feature_importances.sort_values(by='feature_importance', ascending=False).plot(kind='bar', x='features')
plt.ylabel('Feature_importance')
plt.xlabel('')

Text(0.5, 0, '')


# Create the model
model_final = LinearRegression()

# Fit the model on the training data
model_final.fit(x_train_final, y_train_final)

LinearRegression()

LinearRegression()


# Make predictions on the test data
pred_test_final = model_final.predict(x_test_final)


# Create the dummy model
dummy_regr = DummyRegressor(strategy="mean")
# Fit the model on the training data
dummy_regr.fit(x_train_final, y_train_final)
# Make predictions on the test data
pred_dummy_final = dummy_regr.predict(x_test_final)


# Evaluate the model
smape_model_final = smape_score(y_test_final, pred_test_final)
smape_dummy_final = smape_score(y_test_final, pred_dummy_final)

print('sMAPE:', round(smape_model_final, 4))
print('sMAPE случайной модели:', round(smape_dummy_final, 4))

sMAPE: 9.4904
sMAPE случайной модели: 10.1791


f'Итоговая sMAPE: {final_smape(smape_model_rougher, smape_model_final):.2f}'

'Итоговая sMAPE: 9.14'

	date	final.output.concentrate_ag	final.output.concentrate_pb	final.output.concentrate_sol	final.output.concentrate_au	final.output.recovery	final.output.tail_ag	final.output.tail_pb	final.output.tail_sol	final.output.tail_au	primary_cleaner.input.sulfate	primary_cleaner.input.depressant	primary_cleaner.input.feed_size	primary_cleaner.input.xanthate	primary_cleaner.output.concentrate_ag	primary_cleaner.output.concentrate_pb	primary_cleaner.output.concentrate_sol	primary_cleaner.output.concentrate_au	primary_cleaner.output.tail_ag	primary_cleaner.output.tail_pb	primary_cleaner.output.tail_sol	primary_cleaner.output.tail_au	primary_cleaner.state.floatbank8_a_air	primary_cleaner.state.floatbank8_a_level	primary_cleaner.state.floatbank8_b_air	primary_cleaner.state.floatbank8_b_level	primary_cleaner.state.floatbank8_c_air	primary_cleaner.state.floatbank8_c_level	primary_cleaner.state.floatbank8_d_air	primary_cleaner.state.floatbank8_d_level	rougher.calculation.sulfate_to_au_concentrate	rougher.calculation.floatbank10_sulfate_to_au_feed	rougher.calculation.floatbank11_sulfate_to_au_feed	rougher.calculation.au_pb_ratio	rougher.input.feed_ag	rougher.input.feed_pb	rougher.input.feed_rate	rougher.input.feed_size	rougher.input.feed_sol	rougher.input.feed_au	rougher.input.floatbank10_sulfate	rougher.input.floatbank10_xanthate	rougher.input.floatbank11_sulfate	rougher.input.floatbank11_xanthate	rougher.output.concentrate_ag	rougher.output.concentrate_pb	rougher.output.concentrate_sol	rougher.output.concentrate_au	rougher.output.recovery	rougher.output.tail_ag	rougher.output.tail_pb	rougher.output.tail_sol	rougher.output.tail_au	rougher.state.floatbank10_a_air	rougher.state.floatbank10_a_level	rougher.state.floatbank10_b_air	rougher.state.floatbank10_b_level	rougher.state.floatbank10_c_air	rougher.state.floatbank10_c_level	rougher.state.floatbank10_d_air	rougher.state.floatbank10_d_level	rougher.state.floatbank10_e_air	rougher.state.floatbank10_e_level	rougher.state.floatbank10_f_air	rougher.state.floatbank10_f_level	secondary_cleaner.output.tail_ag	secondary_cleaner.output.tail_pb	secondary_cleaner.output.tail_sol	secondary_cleaner.output.tail_au	secondary_cleaner.state.floatbank2_a_air	secondary_cleaner.state.floatbank2_a_level	secondary_cleaner.state.floatbank2_b_air	secondary_cleaner.state.floatbank2_b_level	secondary_cleaner.state.floatbank3_a_air	secondary_cleaner.state.floatbank3_a_level	secondary_cleaner.state.floatbank3_b_air	secondary_cleaner.state.floatbank3_b_level	secondary_cleaner.state.floatbank4_a_air	secondary_cleaner.state.floatbank4_a_level	secondary_cleaner.state.floatbank4_b_air	secondary_cleaner.state.floatbank4_b_level	secondary_cleaner.state.floatbank5_a_air	secondary_cleaner.state.floatbank5_a_level	secondary_cleaner.state.floatbank5_b_air	secondary_cleaner.state.floatbank5_b_level	secondary_cleaner.state.floatbank6_a_air	secondary_cleaner.state.floatbank6_a_level
0	2016-01-15 00:00:00	6.0554	9.8896	5.5073	42.1920	70.5412	10.4120	0.8954	16.9043	2.1431	127.0920	10.1283	7.25	0.9888	8.5476	10.3896	19.5293	34.1744	14.9365	2.5349	7.4761	2.1067	1549.7758	-498.9121	1551.4342	-516.4034	1549.8739	-498.6666	1554.3674	-493.4281	41885.7070	3481.7791	3520.3372	2.8387	6.1004	2.2849	523.5463	55.4866	36.8086	6.4861	11.9866	6.0080	11.8367	6.0058	11.5008	7.1011	28.0293	19.7938	87.1078	5.0080	0.5087	19.1543	1.1702	999.7069	-404.0670	1603.0114	-434.7150	1602.3750	-442.2045	1598.9373	-451.2941	1404.4720	-455.4630	1416.3550	-451.9396	14.5002	4.6948	8.7646	2.6062	25.8531	-498.5265	23.8937	-501.4063	23.9618	-495.2628	21.9404	-499.3410	14.0168	-502.4880	12.0999	-504.7159	9.9256	-498.3102	8.0797	-500.4710	14.1513	-605.8420
1	2016-01-15 01:00:00	6.0294	9.9689	5.2578	42.7016	69.2662	10.4627	0.9275	16.6345	2.2249	125.6292	10.2963	7.25	1.0027	8.5587	10.4971	19.3691	34.1185	16.2505	3.0496	6.7339	2.3530	1576.1667	-500.9050	1575.9506	-499.8659	1575.9942	-499.3151	1574.4793	-498.9317	42050.8618	3498.3710	3489.9819	2.8590	6.1611	2.2660	525.2906	57.2787	35.7534	6.4786	11.9712	6.0058	11.9962	6.0126	11.6159	7.2788	28.0671	20.0510	86.8433	4.9554	0.5367	18.9652	1.1848	1000.2864	-400.0652	1600.7546	-449.9534	1600.4796	-449.8306	1600.5276	-449.9536	1399.2271	-450.8698	1399.7195	-450.1190	14.2655	4.5925	9.0015	2.4882	25.8805	-499.9897	23.8895	-500.3724	23.9706	-500.0855	22.0857	-499.4469	13.9923	-505.5033	11.9505	-501.3315	10.0392	-500.1700	7.9848	-500.5822	13.9984	-599.7872
2	2016-01-15 02:00:00	6.0559	10.2140	5.3838	42.6575	68.1164	10.5070	0.9537	16.2088	2.2579	123.8198	11.3163	7.25	0.9913	8.6035	10.3545	19.1676	33.9695	16.4918	3.1247	6.4718	2.4168	1601.5562	-499.9978	1600.3867	-500.6078	1602.0035	-500.8701	1599.5415	-499.8274	42018.1012	3495.3489	3502.3598	2.9460	6.1165	2.1596	530.0266	57.5106	35.9716	6.3622	11.9206	6.1974	11.9203	6.2046	11.6958	7.2168	27.4540	19.7372	86.8423	4.8435	0.5464	18.8085	1.1626	999.7196	-400.0740	1599.3373	-450.0085	1599.6728	-449.9545	1599.8493	-449.9542	1399.1809	-449.9376	1400.3167	-450.5271	14.1157	4.6248	8.8429	2.4582	26.0052	-499.9296	23.8867	-499.9519	23.9135	-499.4423	23.9577	-499.9020	14.0150	-502.5209	11.9128	-501.1334	10.0709	-500.1291	8.0139	-500.5176	14.0287	-601.4274
3	2016-01-15 03:00:00	6.0480	9.9770	4.8586	42.6898	68.3475	10.4228	0.8838	16.5328	2.1468	122.2702	11.3221	7.25	0.9967	7.2219	8.4966	15.9785	28.2607	16.0244	2.9604	6.8438	2.2621	1599.9687	-500.9518	1600.6592	-499.6771	1600.3041	-500.7280	1600.4495	-500.0526	42029.4480	3498.5783	3499.1629	3.0023	6.0433	2.0378	542.5904	57.7927	36.8622	6.1182	11.6301	6.2032	11.6924	6.1966	11.9150	7.1756	27.3413	19.3208	87.2264	4.6553	0.5425	19.3302	1.0798	999.8148	-400.2002	1600.0594	-450.6199	1600.0128	-449.9105	1597.7252	-450.1301	1400.9432	-450.0301	1400.2347	-449.7908	13.7321	4.4825	9.1229	2.3221	25.9425	-499.1767	23.9555	-499.8488	23.9668	-500.0088	23.9544	-499.9447	14.0365	-500.8573	11.9996	-501.1937	9.9704	-499.2016	7.9773	-500.2559	14.0056	-599.9961
4	2016-01-15 04:00:00	6.1486	10.1425	4.9394	42.7741	66.9270	10.3603	0.7928	16.5257	2.0553	117.9882	11.9136	7.25	1.0099	9.0894	9.9868	19.1999	33.0449	16.4802	3.1121	6.5502	2.2771	1601.3397	-498.9755	1601.4379	-500.3232	1599.5819	-500.8882	1602.6495	-500.5930	42125.3542	3494.8008	3506.6793	3.1696	6.0609	1.7869	540.5319	56.0472	34.3477	5.6637	10.9578	6.1988	10.9605	6.1949	12.4111	7.2402	27.0410	19.2161	86.6888	4.5528	0.5154	19.2674	1.0126	999.6787	-399.7527	1600.2088	-449.5996	1600.3577	-450.0344	1599.7590	-449.9098	1401.5609	-448.8772	1401.1602	-450.4071	14.0800	4.4707	8.8710	2.3304	26.0248	-500.2791	23.9553	-500.5936	23.9857	-500.0838	23.9589	-499.9903	14.0273	-499.8386	11.9531	-501.0539	9.9257	-501.6867	7.8942	-500.3560	13.9966	-601.4967
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
19434	2018-08-18 06:59:59	3.2249	11.3562	6.8035	46.7140	73.7551	8.7696	3.1415	10.4032	1.5292	123.3818	8.0289	6.50	1.3042	6.5592	12.4460	7.0563	32.9402	22.8546	6.5315	3.9343	2.3144	1648.4212	-400.3822	1648.7420	-400.3597	1648.5782	-399.3636	1648.8340	-399.6692	45912.9426	2497.6530	2499.1780	1.1556	6.0919	4.6176	560.8891	85.7183	37.3698	5.3359	7.7628	9.1586	7.7667	9.1561	11.1249	10.9840	30.0689	18.6036	89.5744	4.2073	0.6129	16.6666	0.7486	1199.2459	-300.8455	1149.8079	-498.7897	1047.9636	-498.4131	946.6410	-499.1525	849.6649	-499.2145	849.7581	-497.4487	0.0000	0.0000	0.0000	0.0000	35.0432	-499.0457	29.9067	-499.9799	26.0024	-499.9534	22.9872	-499.9674	23.0315	-501.1679	20.0076	-499.7400	18.0060	-499.8344	13.0011	-500.1557	20.0078	-501.2964
19435	2018-08-18 07:59:59	3.1960	11.3494	6.8622	46.8668	69.0493	8.8973	3.1305	10.5495	1.6125	120.8782	7.9626	6.50	1.3024	6.6124	12.6239	7.5423	32.9253	23.0025	6.6297	4.0891	2.4629	1649.8202	-399.9310	1649.3575	-399.7212	1648.6562	-401.1958	1649.7251	-400.6363	46200.0996	2614.4039	2518.5501	1.1673	6.1213	4.1450	559.0318	119.4992	38.5916	4.8386	7.3567	9.3050	7.0955	9.2979	11.4260	10.8882	29.7840	18.4414	87.7240	4.1779	0.6506	16.9606	0.7716	1196.5693	-299.5122	1147.6752	-500.6083	1048.5657	-500.9328	949.7736	-500.0231	848.5152	-500.2894	850.0131	-496.8221	0.0000	0.0000	0.0000	0.0000	35.0261	-499.8919	29.9218	-499.9497	26.0317	-500.3846	22.9911	-500.0796	22.9601	-501.6128	20.0357	-500.2514	17.9985	-500.3952	12.9540	-499.8952	19.9685	-501.0416
19436	2018-08-18 08:59:59	3.1100	11.4344	6.8860	46.7957	67.0022	8.5296	2.9114	11.1151	1.5966	105.6661	7.9551	6.50	1.3159	7.1289	12.6336	7.9419	31.8567	22.2981	6.2783	4.4575	2.6122	1649.1668	-399.8886	1649.1969	-399.6776	1647.8970	-399.9883	1649.7727	-399.8319	44585.1813	2510.8135	2510.1424	1.1256	5.9705	4.0200	555.6829	122.2627	40.0740	4.5251	6.5860	9.2996	6.5841	9.3001	8.5235	8.9551	22.7865	15.1112	88.8906	4.0500	0.6361	18.3231	0.6851	1204.8666	-299.2357	1149.9429	-501.7179	1049.6044	-500.5491	952.7027	-502.3523	849.0160	-500.5057	850.4556	-506.8980	0.0000	0.0000	0.0000	0.0000	35.0036	-501.0838	29.9905	-611.8559	25.9484	-500.0673	22.9683	-499.8394	23.0157	-501.7116	19.9512	-499.8570	18.0195	-500.4512	13.0234	-499.9144	19.9909	-501.5185
19437	2018-08-18 09:59:59	3.3672	11.6256	6.7994	46.4082	65.5232	8.7772	2.8192	10.4638	1.6029	98.8805	7.9842	6.50	1.2420	7.4904	12.0124	6.9430	30.7709	22.7147	6.2531	3.5571	2.8006	1646.5478	-398.9771	1648.2122	-400.3833	1648.9174	-399.5213	1651.4986	-399.7453	44084.7357	2507.5725	2507.7234	1.1179	6.0481	3.9025	544.7317	123.7424	39.7139	4.3628	6.2101	9.2977	6.2095	9.2972	11.6588	10.6554	29.0527	17.8348	89.8581	3.9025	0.6256	19.4237	0.5671	1201.9042	-299.9073	1154.0878	-500.0366	1054.0098	-500.2373	944.1388	-496.8670	851.5898	-499.0405	851.3456	-499.1226	0.0000	0.0000	0.0000	0.0000	34.9807	-498.1310	29.9685	-586.0133	25.9717	-499.6084	22.9584	-499.8213	23.0250	-501.1534	20.0541	-500.3147	17.9795	-499.2729	12.9924	-499.9763	20.0140	-500.6255
19438	2018-08-18 10:59:59	3.5984	11.7378	6.7175	46.2994	70.2815	8.4067	2.5175	10.6522	1.3894	95.2484	8.0790	6.50	1.2830	7.5376	11.6503	6.2036	30.3566	22.7359	6.0843	3.0267	2.7080	1648.7599	-399.8621	1650.1354	-399.9573	1648.8319	-400.5861	1649.4646	-400.6733	44082.8666	2487.4778	2488.6261	1.1264	6.1587	3.8757	555.8202	94.5444	39.1351	4.3655	6.1470	9.3086	6.1689	9.3099	11.9595	10.7021	29.0782	17.8041	89.5150	3.9255	0.6387	19.9467	0.5864	1196.2381	-299.8627	1147.2482	-500.3632	1047.2791	-500.3541	948.7566	-498.4394	849.4419	-499.2555	850.1122	-499.4071	0.0000	0.0000	0.0000	0.0000	34.9409	-500.1505	30.0319	-500.3283	26.0340	-500.1478	22.9523	-500.0377	23.0186	-500.4927	20.0202	-500.2203	17.9635	-499.9395	12.9903	-500.0810	19.9903	-499.1916

	date	final.output.concentrate_ag	final.output.concentrate_pb	final.output.concentrate_sol	final.output.concentrate_au	final.output.recovery	final.output.tail_ag	final.output.tail_pb	final.output.tail_sol	final.output.tail_au	primary_cleaner.input.sulfate	primary_cleaner.input.depressant	primary_cleaner.input.feed_size	primary_cleaner.input.xanthate	primary_cleaner.output.concentrate_ag	primary_cleaner.output.concentrate_pb	primary_cleaner.output.concentrate_sol	primary_cleaner.output.concentrate_au	primary_cleaner.output.tail_ag	primary_cleaner.output.tail_pb	primary_cleaner.output.tail_sol	primary_cleaner.output.tail_au	primary_cleaner.state.floatbank8_a_air	primary_cleaner.state.floatbank8_a_level	primary_cleaner.state.floatbank8_b_air	primary_cleaner.state.floatbank8_b_level	primary_cleaner.state.floatbank8_c_air	primary_cleaner.state.floatbank8_c_level	primary_cleaner.state.floatbank8_d_air	primary_cleaner.state.floatbank8_d_level	rougher.calculation.sulfate_to_au_concentrate	rougher.calculation.floatbank10_sulfate_to_au_feed	rougher.calculation.floatbank11_sulfate_to_au_feed	rougher.calculation.au_pb_ratio	rougher.input.feed_ag	rougher.input.feed_pb	rougher.input.feed_rate	rougher.input.feed_size	rougher.input.feed_sol	rougher.input.feed_au	rougher.input.floatbank10_sulfate	rougher.input.floatbank10_xanthate	rougher.input.floatbank11_sulfate	rougher.input.floatbank11_xanthate	rougher.output.concentrate_ag	rougher.output.concentrate_pb	rougher.output.concentrate_sol	rougher.output.concentrate_au	rougher.output.recovery	rougher.output.tail_ag	rougher.output.tail_pb	rougher.output.tail_sol	rougher.output.tail_au	rougher.state.floatbank10_a_air	rougher.state.floatbank10_a_level	rougher.state.floatbank10_b_air	rougher.state.floatbank10_b_level	rougher.state.floatbank10_c_air	rougher.state.floatbank10_c_level	rougher.state.floatbank10_d_air	rougher.state.floatbank10_d_level	rougher.state.floatbank10_e_air	rougher.state.floatbank10_e_level	rougher.state.floatbank10_f_air	rougher.state.floatbank10_f_level	secondary_cleaner.output.tail_ag	secondary_cleaner.output.tail_pb	secondary_cleaner.output.tail_sol	secondary_cleaner.output.tail_au	secondary_cleaner.state.floatbank2_a_air	secondary_cleaner.state.floatbank2_a_level	secondary_cleaner.state.floatbank2_b_air	secondary_cleaner.state.floatbank2_b_level	secondary_cleaner.state.floatbank3_a_air	secondary_cleaner.state.floatbank3_a_level	secondary_cleaner.state.floatbank3_b_air	secondary_cleaner.state.floatbank3_b_level	secondary_cleaner.state.floatbank4_a_air	secondary_cleaner.state.floatbank4_a_level	secondary_cleaner.state.floatbank4_b_air	secondary_cleaner.state.floatbank4_b_level	secondary_cleaner.state.floatbank5_a_air	secondary_cleaner.state.floatbank5_a_level	secondary_cleaner.state.floatbank5_b_air	secondary_cleaner.state.floatbank5_b_level	secondary_cleaner.state.floatbank6_a_air	secondary_cleaner.state.floatbank6_a_level
0	2016-01-15 00:00:00	6.0554	9.8896	5.5073	42.1920	70.5412	10.4120	0.8954	16.9043	2.1431	127.0920	10.1283	7.25	0.9888	8.5476	10.3896	19.5293	34.1744	14.9365	2.5349	7.4761	2.1067	1549.7758	-498.9121	1551.4342	-516.4034	1549.8739	-498.6666	1554.3674	-493.4281	41885.7070	3481.7791	3520.3372	2.8387	6.1004	2.2849	523.5463	55.4866	36.8086	6.4861	11.9866	6.0080	11.8367	6.0058	11.5008	7.1011	28.0293	19.7938	87.1078	5.0080	0.5087	19.1543	1.1702	999.7069	-404.0670	1603.0114	-434.7150	1602.3750	-442.2045	1598.9373	-451.2941	1404.4720	-455.4630	1416.3550	-451.9396	14.5002	4.6948	8.7646	2.6062	25.8531	-498.5265	23.8937	-501.4063	23.9618	-495.2628	21.9404	-499.3410	14.0168	-502.4880	12.0999	-504.7159	9.9256	-498.3102	8.0797	-500.4710	14.1513	-605.8420
1	2016-01-15 01:00:00	6.0294	9.9689	5.2578	42.7016	69.2662	10.4627	0.9275	16.6345	2.2249	125.6292	10.2963	7.25	1.0027	8.5587	10.4971	19.3691	34.1185	16.2505	3.0496	6.7339	2.3530	1576.1667	-500.9050	1575.9506	-499.8659	1575.9942	-499.3151	1574.4793	-498.9317	42050.8618	3498.3710	3489.9819	2.8590	6.1611	2.2660	525.2906	57.2787	35.7534	6.4786	11.9712	6.0058	11.9962	6.0126	11.6159	7.2788	28.0671	20.0510	86.8433	4.9554	0.5367	18.9652	1.1848	1000.2864	-400.0652	1600.7546	-449.9534	1600.4796	-449.8306	1600.5276	-449.9536	1399.2271	-450.8698	1399.7195	-450.1190	14.2655	4.5925	9.0015	2.4882	25.8805	-499.9897	23.8895	-500.3724	23.9706	-500.0855	22.0857	-499.4469	13.9923	-505.5033	11.9505	-501.3315	10.0392	-500.1700	7.9848	-500.5822	13.9984	-599.7872
2	2016-01-15 02:00:00	6.0559	10.2140	5.3838	42.6575	68.1164	10.5070	0.9537	16.2088	2.2579	123.8198	11.3163	7.25	0.9913	8.6035	10.3545	19.1676	33.9695	16.4918	3.1247	6.4718	2.4168	1601.5562	-499.9978	1600.3867	-500.6078	1602.0035	-500.8701	1599.5415	-499.8274	42018.1012	3495.3489	3502.3598	2.9460	6.1165	2.1596	530.0266	57.5106	35.9716	6.3622	11.9206	6.1974	11.9203	6.2046	11.6958	7.2168	27.4540	19.7372	86.8423	4.8435	0.5464	18.8085	1.1626	999.7196	-400.0740	1599.3373	-450.0085	1599.6728	-449.9545	1599.8493	-449.9542	1399.1809	-449.9376	1400.3167	-450.5271	14.1157	4.6248	8.8429	2.4582	26.0052	-499.9296	23.8867	-499.9519	23.9135	-499.4423	23.9577	-499.9020	14.0150	-502.5209	11.9128	-501.1334	10.0709	-500.1291	8.0139	-500.5176	14.0287	-601.4274
3	2016-01-15 03:00:00	6.0480	9.9770	4.8586	42.6898	68.3475	10.4228	0.8838	16.5328	2.1468	122.2702	11.3221	7.25	0.9967	7.2219	8.4966	15.9785	28.2607	16.0244	2.9604	6.8438	2.2621	1599.9687	-500.9518	1600.6592	-499.6771	1600.3041	-500.7280	1600.4495	-500.0526	42029.4480	3498.5783	3499.1629	3.0023	6.0433	2.0378	542.5904	57.7927	36.8622	6.1182	11.6301	6.2032	11.6924	6.1966	11.9150	7.1756	27.3413	19.3208	87.2264	4.6553	0.5425	19.3302	1.0798	999.8148	-400.2002	1600.0594	-450.6199	1600.0128	-449.9105	1597.7252	-450.1301	1400.9432	-450.0301	1400.2347	-449.7908	13.7321	4.4825	9.1229	2.3221	25.9425	-499.1767	23.9555	-499.8488	23.9668	-500.0088	23.9544	-499.9447	14.0365	-500.8573	11.9996	-501.1937	9.9704	-499.2016	7.9773	-500.2559	14.0056	-599.9961
4	2016-01-15 04:00:00	6.1486	10.1425	4.9394	42.7741	66.9270	10.3603	0.7928	16.5257	2.0553	117.9882	11.9136	7.25	1.0099	9.0894	9.9868	19.1999	33.0449	16.4802	3.1121	6.5502	2.2771	1601.3397	-498.9755	1601.4379	-500.3232	1599.5819	-500.8882	1602.6495	-500.5930	42125.3542	3494.8008	3506.6793	3.1696	6.0609	1.7869	540.5319	56.0472	34.3477	5.6637	10.9578	6.1988	10.9605	6.1949	12.4111	7.2402	27.0410	19.2161	86.6888	4.5528	0.5154	19.2674	1.0126	999.6787	-399.7527	1600.2088	-449.5996	1600.3577	-450.0344	1599.7590	-449.9098	1401.5609	-448.8772	1401.1602	-450.4071	14.0800	4.4707	8.8710	2.3304	26.0248	-500.2791	23.9553	-500.5936	23.9857	-500.0838	23.9589	-499.9903	14.0273	-499.8386	11.9531	-501.0539	9.9257	-501.6867	7.8942	-500.3560	13.9966	-601.4967
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
14144	2018-08-18 06:59:59	3.2249	11.3562	6.8035	46.7140	73.7551	8.7696	3.1415	10.4032	1.5292	123.3818	8.0289	6.50	1.3042	6.5592	12.4460	7.0563	32.9402	22.8546	6.5315	3.9343	2.3144	1648.4212	-400.3822	1648.7420	-400.3597	1648.5782	-399.3636	1648.8340	-399.6692	45912.9426	2497.6530	2499.1780	1.1556	6.0919	4.6176	560.8891	85.7183	37.3698	5.3359	7.7628	9.1586	7.7667	9.1561	11.1249	10.9840	30.0689	18.6036	89.5744	4.2073	0.6129	16.6666	0.7486	1199.2459	-300.8455	1149.8079	-498.7897	1047.9636	-498.4131	946.6410	-499.1525	849.6649	-499.2145	849.7581	-497.4487	0.0000	0.0000	0.0000	0.0000	35.0432	-499.0457	29.9067	-499.9799	26.0024	-499.9534	22.9872	-499.9674	23.0315	-501.1679	20.0076	-499.7400	18.0060	-499.8344	13.0011	-500.1557	20.0078	-501.2964
14145	2018-08-18 07:59:59	3.1960	11.3494	6.8622	46.8668	69.0493	8.8973	3.1305	10.5495	1.6125	120.8782	7.9626	6.50	1.3024	6.6124	12.6239	7.5423	32.9253	23.0025	6.6297	4.0891	2.4629	1649.8202	-399.9310	1649.3575	-399.7212	1648.6562	-401.1958	1649.7251	-400.6363	46200.0996	2614.4039	2518.5501	1.1673	6.1213	4.1450	559.0318	119.4992	38.5916	4.8386	7.3567	9.3050	7.0955	9.2979	11.4260	10.8882	29.7840	18.4414	87.7240	4.1779	0.6506	16.9606	0.7716	1196.5693	-299.5122	1147.6752	-500.6083	1048.5657	-500.9328	949.7736	-500.0231	848.5152	-500.2894	850.0131	-496.8221	0.0000	0.0000	0.0000	0.0000	35.0261	-499.8919	29.9218	-499.9497	26.0317	-500.3846	22.9911	-500.0796	22.9601	-501.6128	20.0357	-500.2514	17.9985	-500.3952	12.9540	-499.8952	19.9685	-501.0416
14146	2018-08-18 08:59:59	3.1100	11.4344	6.8860	46.7957	67.0022	8.5296	2.9114	11.1151	1.5966	105.6661	7.9551	6.50	1.3159	7.1289	12.6336	7.9419	31.8567	22.2981	6.2783	4.4575	2.6122	1649.1668	-399.8886	1649.1969	-399.6776	1647.8970	-399.9883	1649.7727	-399.8319	44585.1813	2510.8135	2510.1424	1.1256	5.9705	4.0200	555.6829	122.2627	40.0740	4.5251	6.5860	9.2996	6.5841	9.3001	8.5235	8.9551	22.7865	15.1112	88.8906	4.0500	0.6361	18.3231	0.6851	1204.8666	-299.2357	1149.9429	-501.7179	1049.6044	-500.5491	952.7027	-502.3523	849.0160	-500.5057	850.4556	-506.8980	0.0000	0.0000	0.0000	0.0000	35.0036	-501.0838	29.9905	-611.8559	25.9484	-500.0673	22.9683	-499.8394	23.0157	-501.7116	19.9512	-499.8570	18.0195	-500.4512	13.0234	-499.9144	19.9909	-501.5185
14147	2018-08-18 09:59:59	3.3672	11.6256	6.7994	46.4082	65.5232	8.7772	2.8192	10.4638	1.6029	98.8805	7.9842	6.50	1.2420	7.4904	12.0124	6.9430	30.7709	22.7147	6.2531	3.5571	2.8006	1646.5478	-398.9771	1648.2122	-400.3833	1648.9174	-399.5213	1651.4986	-399.7453	44084.7357	2507.5725	2507.7234	1.1179	6.0481	3.9025	544.7317	123.7424	39.7139	4.3628	6.2101	9.2977	6.2095	9.2972	11.6588	10.6554	29.0527	17.8348	89.8581	3.9025	0.6256	19.4237	0.5671	1201.9042	-299.9073	1154.0878	-500.0366	1054.0098	-500.2373	944.1388	-496.8670	851.5898	-499.0405	851.3456	-499.1226	0.0000	0.0000	0.0000	0.0000	34.9807	-498.1310	29.9685	-586.0133	25.9717	-499.6084	22.9584	-499.8213	23.0250	-501.1534	20.0541	-500.3147	17.9795	-499.2729	12.9924	-499.9763	20.0140	-500.6255
14148	2018-08-18 10:59:59	3.5984	11.7378	6.7175	46.2994	70.2815	8.4067	2.5175	10.6522	1.3894	95.2484	8.0790	6.50	1.2830	7.5376	11.6503	6.2036	30.3566	22.7359	6.0843	3.0267	2.7080	1648.7599	-399.8621	1650.1354	-399.9573	1648.8319	-400.5861	1649.4646	-400.6733	44082.8666	2487.4778	2488.6261	1.1264	6.1587	3.8757	555.8202	94.5444	39.1351	4.3655	6.1470	9.3086	6.1689	9.3099	11.9595	10.7021	29.0782	17.8041	89.5150	3.9255	0.6387	19.9467	0.5864	1196.2381	-299.8627	1147.2482	-500.3632	1047.2791	-500.3541	948.7566	-498.4394	849.4419	-499.2555	850.1122	-499.4071	0.0000	0.0000	0.0000	0.0000	34.9409	-500.1505	30.0319	-500.3283	26.0340	-500.1478	22.9523	-500.0377	23.0186	-500.4927	20.0202	-500.2203	17.9635	-499.9395	12.9903	-500.0810	19.9903	-499.1916

	date	primary_cleaner.input.sulfate	primary_cleaner.input.depressant	primary_cleaner.input.feed_size	primary_cleaner.input.xanthate	primary_cleaner.state.floatbank8_a_air	primary_cleaner.state.floatbank8_a_level	primary_cleaner.state.floatbank8_b_air	primary_cleaner.state.floatbank8_b_level	primary_cleaner.state.floatbank8_c_air	primary_cleaner.state.floatbank8_c_level	primary_cleaner.state.floatbank8_d_air	primary_cleaner.state.floatbank8_d_level	rougher.input.feed_ag	rougher.input.feed_pb	rougher.input.feed_rate	rougher.input.feed_size	rougher.input.feed_sol	rougher.input.feed_au	rougher.input.floatbank10_sulfate	rougher.input.floatbank10_xanthate	rougher.input.floatbank11_sulfate	rougher.input.floatbank11_xanthate	rougher.state.floatbank10_a_air	rougher.state.floatbank10_a_level	rougher.state.floatbank10_b_air	rougher.state.floatbank10_b_level	rougher.state.floatbank10_c_air	rougher.state.floatbank10_c_level	rougher.state.floatbank10_d_air	rougher.state.floatbank10_d_level	rougher.state.floatbank10_e_air	rougher.state.floatbank10_e_level	rougher.state.floatbank10_f_air	rougher.state.floatbank10_f_level	secondary_cleaner.state.floatbank2_a_air	secondary_cleaner.state.floatbank2_a_level	secondary_cleaner.state.floatbank2_b_air	secondary_cleaner.state.floatbank2_b_level	secondary_cleaner.state.floatbank3_a_air	secondary_cleaner.state.floatbank3_a_level	secondary_cleaner.state.floatbank3_b_air	secondary_cleaner.state.floatbank3_b_level	secondary_cleaner.state.floatbank4_a_air	secondary_cleaner.state.floatbank4_a_level	secondary_cleaner.state.floatbank4_b_air	secondary_cleaner.state.floatbank4_b_level	secondary_cleaner.state.floatbank5_a_air	secondary_cleaner.state.floatbank5_a_level	secondary_cleaner.state.floatbank5_b_air	secondary_cleaner.state.floatbank5_b_level	secondary_cleaner.state.floatbank6_a_air	secondary_cleaner.state.floatbank6_a_level
0	2016-09-01 00:59:59	210.8009	14.9931	8.0800	1.0050	1398.9813	-500.2256	1399.1449	-499.9197	1400.1030	-500.7044	1399.0017	-499.4851	13.1286	5.6367	489.7937	62.7105	42.0224	12.0841	16.9229	6.1535	16.8675	6.1507	1001.8493	-350.3013	1249.7386	-399.1075	1249.7505	-399.3967	1198.2879	-399.4886	999.4723	-399.5310	949.5662	-398.1808	24.9382	-500.4910	14.9476	-500.0138	20.0182	-450.3960	13.9871	-449.8315	12.0236	-497.7958	8.0167	-501.2891	7.9466	-432.3178	4.8725	-500.0374	26.7059	-499.7094
1	2016-09-01 01:59:59	215.3925	14.9875	8.0800	0.9905	1398.7779	-500.0574	1398.0554	-499.7782	1396.1510	-499.2402	1399.5081	-500.4163	13.0363	5.5261	490.1045	61.9612	41.1879	11.9190	17.0025	5.9990	16.9961	6.0023	998.6908	-350.4257	1248.3955	-399.9460	1249.5145	-399.6308	1200.5065	-399.9424	1000.0021	-399.4915	950.2000	-405.7875	24.9235	-499.8140	14.9309	-500.7645	19.9885	-450.1140	14.0930	-450.0592	12.0581	-498.6958	8.1310	-499.6342	7.9583	-525.8396	4.8789	-500.1624	25.0199	-499.8194
2	2016-09-01 02:59:59	215.2599	12.8849	7.7867	0.9960	1398.4937	-500.8684	1398.8604	-499.7645	1398.0757	-502.1515	1399.4971	-499.7155	13.1384	5.4266	489.6180	66.9038	42.5463	12.0913	16.9925	5.8506	16.9821	5.8536	998.5167	-349.7835	1247.4407	-400.2643	1248.2072	-401.0744	1199.7700	-400.7906	999.9255	-399.2370	950.3205	-400.8643	24.9087	-500.3040	14.9965	-500.9938	20.0397	-450.2632	14.0779	-449.6610	11.9624	-498.7675	8.0969	-500.8274	8.0711	-500.8017	4.9051	-499.8285	24.9949	-500.6226
3	2016-09-01 03:59:59	215.3362	12.0068	7.6400	0.8635	1399.6181	-498.8636	1397.4401	-499.2110	1400.1293	-498.3559	1401.0652	-501.0387	12.4005	5.1148	476.6181	59.8659	41.0602	12.1816	16.5321	5.8019	16.5153	5.8037	1000.2766	-350.1726	1251.3227	-398.6553	1250.4937	-399.7467	1199.3987	-397.5031	1001.9307	-400.4376	950.7353	-399.8030	24.8940	-499.3824	14.9167	-499.8623	20.0311	-449.3726	14.0140	-449.5269	12.0331	-498.3509	8.0749	-499.4744	7.8971	-500.8685	4.9314	-499.9636	24.9489	-498.7100
4	2016-09-01 04:59:59	199.0993	10.6825	7.5300	0.8056	1401.2681	-500.8083	1398.1288	-499.5045	1402.1722	-500.8106	1399.4805	-499.3741	11.3274	4.7674	488.2481	63.3153	41.2692	11.2904	13.6073	5.7381	13.6501	5.7396	996.5413	-350.5625	1304.6588	-399.5115	1306.4555	-399.0507	1248.6993	-400.8769	1058.8391	-398.9882	949.6453	-399.2776	24.8869	-499.3574	14.9792	-500.1868	19.9624	-450.6388	14.0092	-450.0223	12.0254	-500.7865	8.0547	-500.3975	8.1079	-509.5267	4.9577	-500.3600	25.0033	-500.8563
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
5285	2017-12-31 19:59:59	173.9578	15.9634	8.0700	0.8967	1401.9306	-499.7288	1401.4414	-499.1934	1399.8103	-499.5991	1400.3348	-499.2435	11.6390	4.6953	521.6852	66.7585	42.0220	10.5545	10.1092	7.2039	10.1108	7.1982	1399.9353	-299.9996	1049.6420	-500.0749	951.9239	-499.6488	852.3374	-498.4644	851.4368	-500.2856	851.4613	-499.7356	21.9121	-499.9152	19.9091	-499.7349	24.9396	-500.0985	12.9684	-499.9520	13.9960	-500.1575	12.0692	-499.6733	7.9773	-499.5161	5.9333	-499.9660	8.9872	-499.7559
5286	2017-12-31 20:59:59	172.9103	16.0026	8.0700	0.8965	1447.0757	-494.7168	1448.8519	-465.9630	1443.8904	-503.5877	1447.6534	-480.4038	11.3381	4.4784	530.1958	62.5607	41.0964	10.2541	9.9795	7.2009	9.9789	7.2020	1406.1223	-299.0572	1055.7152	-498.2502	956.8503	-497.4039	855.3867	-495.7027	852.0713	-499.4428	851.3290	-496.5798	25.3223	-497.0350	23.3106	-498.8926	26.3692	-499.2224	17.7049	-499.5457	16.7498	-496.0315	13.3654	-499.1227	9.2886	-496.8930	7.3729	-499.9430	8.9868	-499.9038
5287	2017-12-31 21:59:59	171.1357	15.9937	8.0700	1.1660	1498.8362	-501.7704	1499.5724	-495.5163	1502.7492	-520.6674	1502.5853	-498.2602	11.2859	4.4053	546.3069	73.6475	41.0955	10.1691	10.1822	7.1963	10.1855	7.1987	1396.2031	-300.2107	1047.8797	-499.6042	948.9634	-499.5496	856.3256	-500.0956	850.6146	-499.4737	850.2796	-498.5084	29.0759	-500.5284	26.8880	-500.9758	28.0002	-500.5659	22.9965	-500.4776	19.9941	-499.7913	15.1014	-499.9363	10.9892	-498.3479	9.0209	-500.0404	8.9820	-497.7899
5288	2017-12-31 22:59:59	179.6972	15.4390	8.0700	1.5011	1498.4662	-500.4840	1497.9870	-519.2003	1496.5690	-487.4796	1502.6889	-501.5697	11.2897	4.3435	547.3314	78.0664	41.7157	10.0712	10.7309	7.2001	10.7263	7.2008	1403.9423	-300.0301	1049.8023	-499.7956	950.7694	-500.3746	847.6472	-499.9203	849.4130	-500.3805	850.0750	-500.4782	29.1032	-499.6927	27.0517	-499.8583	27.9804	-499.9072	23.0466	-499.7449	19.9588	-499.9587	15.0269	-499.7231	11.0116	-499.9850	9.0098	-499.9379	9.0127	-500.1543
5289	2017-12-31 23:59:59	181.5569	14.9958	8.0700	1.6235	1498.0963	-499.7969	1501.7438	-505.1469	1499.5360	-492.4282	1499.6743	-499.7667	10.9046	4.0906	564.5404	77.6787	40.8911	9.6973	10.7895	7.1958	10.7929	7.2010	1400.6115	-299.9894	1052.7828	-499.7630	950.7937	-499.3572	850.6395	-500.4338	850.3708	-500.8194	850.6802	-498.5392	29.0981	-500.4299	26.8924	-500.0871	28.0235	-500.1704	23.0136	-500.2188	20.0347	-500.7286	14.9142	-499.9485	10.9866	-500.6580	8.9895	-500.3376	8.9886	-500.7649

	rougher.output.recovery	final.output.recovery
count	19439.0000	19439.0000
mean	83.3286	67.0502
std	14.1481	10.1258
min	0.0000	0.0000
25%	81.0380	63.2997
50%	86.1889	68.1727
75%	90.0107	72.6866
max	100.0000	100.0000

	rougher.output.concentrate_au	rougher.output.tail_au	final.output.concentrate_au	final.output.tail_au
count	19439.0000	19439.0000	19439.0000	19439.0000
mean	19.7720	1.8164	44.0765	3.0425
std	3.7536	0.6800	5.1298	0.9228
min	0.0000	0.0207	0.0000	0.0000
25%	18.8025	1.4042	43.4022	2.4611
50%	20.2802	1.8082	45.0112	2.9849
75%	21.7178	2.2119	46.2753	3.5714
max	28.8245	9.6890	52.7566	8.2450

Прогнозирование восстановления золота из руды¶

Содержание

Обзор и подготовка данных¶

Проверка рассчетов¶

Исследовательский анализ данных¶

Анализ признаков и их взаимодействий¶

Целевые признаки и доли золота в разных субстратах¶

Размер гранул сырья и скорость их подачи¶

Флотационные реагенты¶

Физические параметры¶

Доля металлов в руде¶

Изменение концентрации металлов на различных этапах¶

Обработка аномальных значений¶

Обработка пропущенных значений¶

Исследование моделей¶

Прогноз коэффициента обогащения чернового концентрата¶

Линейная регрессия¶

Случайный лес¶

Cлучайная модель¶

Прогноз коэффициента обогащения финального продукта¶

Линейная регрессия¶

Случайный лес¶

Cлучайная модель¶

Тестирование моделей¶

Предсказание коэффициента обогащения для чернового концентрата¶

Предсказание коэффициента обогащения для финального продукта¶

Итоговая метрика¶

Общий вывод:¶

	rougher.input.feed_au	rougher.output.concentrate_au	rougher.output.tail_au	rougher.output.recovery
45	7.1141	0.0	1.9398	0.0
46	7.6515	0.0	2.0876	0.0
47	5.5877	0.0	2.1265	0.0
48	4.7530	0.0	1.9389	0.0
49	5.1228	0.0	2.0509	0.0
...	...	...	...	...
15620	7.9511	0.0	0.7419	0.0
15621	8.7158	0.0	0.8280	0.0
15622	9.0261	0.0	1.0514	0.0
15623	8.7524	0.0	1.0643	0.0
15624	8.4676	0.0	1.0667	0.0

	rougher.input.feed_au	final.output.concentrate_au	final.output.tail_au	rougher.output.recovery	final.output.recovery
644	5.5990	0.0	1.7791	68.9610	0.0
1193	7.7861	0.0	2.9845	82.7873	0.0
1194	7.4992	0.0	2.7473	83.5153	0.0
1195	7.0323	0.0	2.7914	83.7022	0.0
1196	7.0216	0.0	2.6821	84.8368	0.0
...	...	...	...	...	...
14951	11.6048	0.0	5.0618	79.7052	0.0
14952	11.7636	0.0	4.2916	83.3258	0.0
15183	7.3455	0.0	5.8195	26.9599	0.0
15184	7.6878	0.0	3.8373	78.7868	0.0
15310	6.0094	0.0	1.2109	65.0039	0.0

	rougher.input.feed_au	rougher.output.concentrate_au	rougher.output.tail_au	rougher.output.recovery
16226	0.01	0.01	1.3525	100.0
16227	0.01	0.01	1.3525	100.0
16228	0.01	0.01	1.3525	100.0
16229	0.01	0.01	1.3525	100.0
16230	0.01	0.01	1.3525	100.0

	rougher.output.concentrate_au	primary_cleaner.output.concentrate_au	final.output.concentrate_au	final.output.tail_au	final.output.recovery
75	18.8615	5.9645	46.5433	0.0	100.0
1916	22.0130	10.3694	25.7735	0.0	100.0
1927	19.0389	34.1149	42.9840	0.0	100.0
2017	24.8120	28.5552	38.4562	0.0	100.0
3388	16.9222	25.1345	42.7182	0.0	100.0
...	...	...	...	...	...
18326	17.4802	35.1516	46.5568	0.0	100.0
18420	17.2792	28.7369	40.5272	0.0	100.0
19082	21.3183	35.4653	32.2661	0.0	100.0
19209	24.4854	30.8070	1.0640	0.0	100.0
19210	22.5045	32.5089	41.1188	0.0	100.0

	rougher.input.feed_size	rougher.input.feed_rate	primary_cleaner.input.feed_size	rougher.output.recovery	final.output.recovery
count	19294.0000	19428.0000	19439.0000	19439.0000	19439.0000
mean	58.9703	478.3200	7.3071	83.3286	67.0502
std	21.6345	105.3738	0.6135	14.1481	10.1258
min	0.0464	0.0030	1.0800	0.0000	0.0000
25%	47.4351	416.5255	6.9400	81.0380	63.2997
50%	54.6102	499.4161	7.2800	86.1889	68.1727
75%	65.0230	550.1722	7.6700	90.0107	72.6866
max	484.9675	717.5088	15.5000	100.0000	100.0000

	rougher.input.feed_au	rougher.input.feed_ag	rougher.input.feed_pb	rougher.input.feed_sol
count	19439.0000	19439.0000	19339.0000	19340.0000
mean	8.2661	8.7949	3.5981	36.6951
std	1.9551	1.9368	1.0495	5.1454
min	0.0100	0.0100	0.0100	0.0100
25%	6.8544	7.2380	2.8873	34.2078
50%	8.1258	8.5936	3.5349	37.2026
75%	9.7699	10.2138	4.2442	40.0391
max	13.7313	14.5960	7.1426	53.4777